Errors in JACET 8000 Level Marker

I have been using the JACET8000 Level Marker (new link here) page for my research. It is a great tool. And I would like to thank the creator, Shinichi Shimizu, for it. However, when I checked the output I had found a number of errors. Below is a list of words which have been incorrectly numbered. The digit to the left (in red) of the word is what the output gives. The digit to the right (in green) is what the number or level should be:

1 march

2

1 lower

3

1 saw

3

1 means

4

1 thanks

4

1 leading

4

1 finding

4

1 thinking

4

1 china

4

1 sin

4

1 colored

4

1 saying

5

1 forward

5

1 basin

5

1 doing

5

1 making

5

1 controlled

5

1 kin

6

1 mar

8

2 audience

1

2 preferred

6

2 flatter

7

3 interpret

2

3 clothes

4

3 handicapped

6

4 including

3

4 upward

7

5 boom

4

5 ethics

6

5 summons

6

6 constructive

5

6 ragged

7

6 robin

7

7 chatter

6

8 coastline

7

9 alight

8

I had checked this with the March 2003 data as well as the publication “Daigaku Eigo Kyoiku Gakkai Kihongo ni Motozuku JACET 8000 Eitango” (ISBN 9784342788734). This information was correct at the time of posting.

[update] Checking through some of the words I found they were difficult to classify. For example, ‘saw’ could be the past tense of the verb ‘see’. In that case it would be placed in JACET 1. But as a noun – an tool to cut things in half – it would be in JACET 3. But others were clearly errors. ‘Audience’, ‘interpret’, ‘including’, ‘boom’, ‘constructive’, ‘chatter’ and ‘alight’ were numbering errors at the cut-off point between levels. The rest of the errors I cannot figure out.

KH Coder

kh coder logo
KH Coder is a corpus linguistic tool for Japanese texts. Follow this link to learn more about Japanese corpus linguistics.

A Wordsmith Tools discussion list

wordsmith tools 5.0
Mike Scott, the creator of Wordsmith Tools has a mailing list by that name. The archive is open to public, but to post one must become a member of the list, pending approval.

Two useful pages for Range program

Here are two useful pages I found about Paul Nation’s Range program:
BNC spoken freq
TV as language teacher

Fixing the ‘Range check error’ in Range

I finally got Paul Nation’s Range program to work with my base word files, thanks to S.F..

The problem was in the fact I had unnecessary tabs and carriage returns between word families. But not having a digit after each word (like the original files) didn’t seem to affect the program’s performance.